A variety of network management platforms exist to assist network administrators with installing and configuring network resources. In many platforms, a management server can be used to issue commands to hosts or other network nodes to manage the configuration of the network hosts, underlying clients or other devices. In the case of comparatively large-scale networks, and/or networks in which nodes are relatively widely dispersed geographically, the distribution of commands can become more problematic. For one thing, the systems administrator may not be aware of the best connection route or pathway through the network to a set of target machines, to most effectively “push” the commands to their destination. For example, certain systems may be reachable only through intermediate systems due to firewall configurations, and multiple levels of firewalls may be in place. For further example, it may be desirable to reduce or minimize the number of network nodes or “hops” that the commands need to traverse to arrive at the intended target machines. For yet further example, in the case of relatively large-scale networks, for example on the order of hundreds or thousands of hosts and/or nodes or more, it may be desirable to transmit the commands as few times as necessary to avoid repetitively transmitting the commands over again each time. It may thus be desirable to provide methods and systems that overcome these network management difficulties, and permit efficient network management over dispersed networks.